Echinoids from the Tesero Member (Werfen Formation) of the Dolomites (Italy): implications for extinction and survival of echinoids in the aftermath of the end-Permian mass extinction

Geological and stratigraphical setting

Since the 19th century, the Dolomites (northern Italy; Fig. 1) have represented a reference area to study the EPME of the shallow marine biota. The Upper Permian is represented by the Bellerophon Formation, a thick sulphate-carbonate succession, which records marginal (sabkha) to shallow marine conditions (e.g., Massari & Neri, 1997). The upper part of the formation is characterised by black bioclastic limestone, which contains rich benthic assemblages dominated by calcareous algae, foraminifers, molluscs and brachiopods. The Bellerophon Formation is overlain by the Werfen Formation, which begins with the Tesero Member, a transgressive carbonate unit mostly composed of oolitic grainstones, mudstone and microbial limestone (e.g., Bosellini, 1964; Noé, 1987; Beretta et al., 1999; Farabegoli, Perri & Posenato, 2007; Groves et al., 2007; Jia et al., 2012). The basal centimetres of the Tesero Member record the peak of the EPME, while the PTB has been recognized a few decimetres above (e.g., 1.3 m above the formational boundary in the Bulla section; for comprehensive descriptions of the PTB succession of the Dolomites see Farabegoli, Perri & Posenato (2007), Posenato (2009) and Posenato (2010)).

Conodonts are rare in the Tesero section (e.g., Farabegoli, Perri & Posenato, 2007). Hindeodus changxingensis, the marker of the last Permian biozone and proxy of the PTB (Metcalfe, Nicoll & Wardlaw, 2007), occurs 1.3 m above the Bellerophon–Werfen formational boundary (BWFB), while Hindeodus parvus, the marker of the PTB, appears here only at 11 m above the BWFB, after a 9 m interval lacking conodonts. For this reason, the H. parvus first occurrence in the Tesero section has been considered diachronous and younger than that of the Bulla section, another key section for studies of the PTB in the Dolomites. At Bulla, it appears at 1.3 m above the BWFB (Perri & Farabegoli, 2003). Following Posenato (2009) and Farabegoli & Perri (2012), the PTB in the Tesero section is tentatively placed in the upper part of unit 5b, slightly predating bed CNT11A (Fig. 2), where the peak of the articulate brachiopod Teserina nerii occurs. Therefore, the echinoid bearing beds of the Tesero section probably have an age encompassing the PTB.

The Tesero section, easily accessible because of its location along the road connecting the village of Tesero with the main road along the bottom of the Avisio Valley (Fig. 1), is among the most studied PTB successions in the Dolomites. It was first described by Leonardi (1935), and carefully analysed over the course of the last few decades from palaeontological, sedimentological and geochemical points of view (e.g., Bosellini, 1964; Neri & Pasini, 1985; Noé, 1987; Broglio Loriga et al., 1988; Magaritz et al., 1988; Wignall & Hallam, 1992; Beretta et al., 1999; Perri & Farabegoli, 2003; Chen et al., 2006; Groves et al., 2007; Posenato, 2009; Spina et al., 2015; Foster et al., 2017). The section became famous globally after the discovery of a “mixed brachiopod and mollusc fauna” located about 1.5–2.8 m above the BWFB, formerly considered to be coincident with the PTB (Neri & Pasini, 1985).

At Tesero, the black bioclastic limestone of the Bellerophon Fm (Bulla Member) is only about 50 cm thick (Fig. 2, unit 3). It overlies light-grey, marly dolostone representing peritidal environments with root traces and mud-cracks, recording short-term subaerial exposure, in the underlying Bellerophon Formation unit (unit 2). The Tesero Member begins with a basal oolitic bedset, about 90 cm thick (units 4a, 4b), the middle of which records the extinction peak of lagenide foraminifers (Groves et al., 2007). It is followed by an interval of ca. 180 cm (units 5a, 5b) characterised by wide lateral facies variability, composed of oolitic and bioclastic grainstone/packstone, marly interlayers and lenses, mudstone and microbialite dm-size mounds (Beretta et al., 1999, fig. 4). The marlstone is dominated by the articulate brachiopods Orbicoelia dolomitensis Chen and Teserina nerii Posenato (Broglio Loriga et al., 1988; Chen et al., 2006; Posenato, 2009; Posenato, 2011), which are respectively abundant in the lower and upper “mixed fauna beds”. It is these same “mixed fauna beds” from which the echinoids described herein were collected.

In the upper Palaeozoic of northern Italy, two species of echinoid, Archaeocidaris tirolensis Stache, 1876 and Archaeocidaris ladina Stache, 1877 have been named from the Bellerophon Formation. These named taxa are, however, nomina nuda or indeterminate, respectively (Jackson, 1912) and are based off of disarticulated, fragmentary material. In the Tesero section, echinoid remains are abundant in the black limestone of the Bellerophon Fm (Bulla Mb), but cannot be extracted from the matrix because they are embedded in hard, cemented, limestone. The material studied herein has been collected from unit 5b of the Tesero Member (Fig. 2), where plates and spines are preserved on microbial bed surfaces or as isolated specimens extracted from washed residues obtained from disaggregation and sieving of the marlstones between the microbialites. Higher in the Werfen formation, echinoid spines are recorded within the Siusi Member (late Induan) (Broglio Loriga, Masetti & Neri, 1983, 1983, pl. 46).

Systematic palaeontology

Type genus. Miocidaris Döderlein, 1887

Other genera. Eotiaris Lambert, 1899, Couvelardicidaris Vadet, 1991, Procidaris Pomel, 1883

Genus Eotiaris Lambert, 1899

Type species. Cidaris keyserlingi Geinitz, 1848, from the Wuchiapingian Zechstein of Germany and the United Kingdom.

Other species. Eotiaris guadalupensis Thompson 2017 in Thompson, Petsios & Bottjer, 2017 from the Roadian and Capitanian of west Texas and E. connorsi Kier, 1965 from the Capitanian of West Texas.

Diagnosis. Miocidarid with small test. Interambulacral plates imbricate adapically. Areoles confluent at or below ambitus.

Occurrence. Wuchiapingian of Germany, the U.K. and Roadian and Guadalupian of Texas.

Diagnosis. Eotiaris with only first one or two most adoral areoles confluent. Spines long, thin, straight and cylindrical. Some spines striate and some with small distally facing spinules.

Etymology. After the locality Tesero, where the holotype was collected.

Holotype. The holotype is specimen MPL 8651-1 (Figs. 3A–3D).

Material. In addition to the holotype, a number of disarticulated spines and interambulacral plates are attributed to this taxon. These are MPL 8652, MPL 8659, MPL 8653-1, MPL 8656-2, MPL 8656-1, MPL 8657-5, MPL 8655-1, MPL 8653-5, MPL 8653-3, MPL 8657-4, MPL 8657-3, MPL 8658-5, MPL 8654-2, and MPL 8657-6.

Occurrence. This species is known from the latest Changhsingian –early Induan Tesero Member at the Tesero section, Dolomites of Italy.

Description. Test regular and small. Only a single interambulacral area 6.1 mm in width is known from specimen MPL 8651-1 (Fig. 3), but using the calculation of Smith & Hollingworth (1990), the test diameter would likely have been about 13.5 mm. Hemipyramids and rotulae that may belong to this taxon are described below as Echinoidea, gen. et sp. indet. The peristomial plating, apical plating and details of the perignathic girdle are unknown.

The interambulacrum is composed of two columns of plates (Fig. 3). At and below the ambitus the plating appears to be rigid, while plates above the ambitus are missing. Five plates are preserved in each column, though more were undoubtedly present. These missing adapical plates may have been imbricate as is the case for other species in the genus (e.g., E. keyserlingi). Interambulacral plates are pentagonal, ranging from 1.5 to 2.1 times wider than high (Table 1). Interambulacral tubercles are perforate and crenulate, slightly sunken. The one or two most adoral plates with confluent areoles adorally and adapically (Fig. 3). One disarticulated plate attributed to this taxon without confluent areoles (Fig. 5C) so more adapical plates were likely without confluent areoles. Areoles at the ambitus are circular, from 1.4 to 1.7 mm high and 1.5 to 1.8 mm wide (Table 1). Each plate is covered with numerous imperforate, non-crenulate secondary tubercles, which are arranged in a closest-packing arrangement and cover the entire extrascrobicular area. There are about 40 secondary tubercles on the largest ambital plates, and about 25 on the second most adoral plate (Fig. 3). At the ambitus there are three rows of scrobicular tubercles between the areoles and the adambulacral suture and three rows between the areole and interradial suture. On the second most adoral plates, there are about two rows of scrobicular tubercles between the areole and adambulacral suture and one or two between the areole and interradial suture. Scrobicular tubercles differentiated into slightly distinct scrobicular ring (Figs. 5A, 5C).

Spines attributed to this species are long, thin. They are straight and cylindrical or slightly fusiform. Some spines striate and lacking ornamentation (Figs. 5J, 5L) and some with small distally facing spinules located distally (Figs. 5F–5H). All spines with a well-defined milled ring (Figs. 5E–5L). Straight spines widest at the milled ring, fusiform spines widest at about midway their length (Fig. 5L). Longest spine fragment is 17.4 mm in length, though this is not associated with any test thus the spine length relative to test size is unknown. Most other complete spines are much shorter.

Remarks. The two columned interambulacral area, morphology of spines, confluence of areoles, imbricate adradial suture and imbricate adapical interambulacral plating identify this species as a member of the genus Eotiaris. The specimens described herein are morphologically similar to many specimens of Eotiaris keyserlingi (Geinitz, 1848) known from the Wuchiapingian Zechstein of Germany (Reich, 2007), the United Kingdom (Smith & Hollingworth, 1990) and reported from Poland (Peryt et al., 2016). E. teseroensis differs from E. keyserlingi with respect to the confluence of it’s areoles. While only the first one or two areoles of E. teseroensis are confluent, in E. keyserlingi, about the four most adoral interambulacral plates bear confluent areoles. Other than the features of the interambulacral plates, the main differentiating feature of species of Eotiaris is the morphology of the spines. Both E. keyserlingi from the Zechstein and E. teseroensis described herein have spines which are smooth, taper distally and bear small spinules at their distal end (Compare Figs. 5F, 5H with fig. 4.6 in Smith & Hollingworth, 1990). There apparently exists much variability in the spines attributed to E. keyserlingi, as those shown in Fig. 3C by Reich (2007) bear many more spinules along their proximal shaft than those in fig. 4.6 of Smith & Hollingworth (1990). Furthermore, some spines attributed to E. keyserlingi appear to show no spinules, and are instead smooth for their entire length. Many echinoids, both extant and fossil, show aboral to oral variability in spine morphology along their test (Mortensen, 1928; Schneider, Sprinkle & Ryder, 2005), thus it should be no surprise that this is found in E. keyserlingi. In our washed samples and on bedding planes, we also find variability in spine morphology, from thin striate spines with distal spinules to fusiform spines lacking ornamentation (Figs. 5F–5L). E. teseroensis differs substantially from E. guadalupensis Thompson in Thompson, Petsios & Bottjer, 2017 from the Roadian to Capitanian of west Texas in the morphology of its spines. E. guadalupensis has spines covered with numerous spinules which are straight to fusiform to clavate or bulbous in shape (Thompson et al., 2015; Thompson, Petsios & Bottjer, 2017). The spines attributed to E. teseroensis, however, are much less wide and do not have the large, bulbous or clavate morphology or distinct spinules of E. guadalupensis. E. teseroensis, E. keyserlingi, and E. guadalupensis all differ from E. connorsi Kier, 1965 in having narrower interambulacral plates, denser secondary tuberculation on their interambulacral plates, and more rigidly sutured plating adorally. Phylogenetic analyses also indicate that E. connorsi may be more basal than the other species of Eotiaris (Thompson et al., in press). Despite the fact that the echinoids described herein largely differ morphologically from E. keyserlingi only with respect to the number of confluent areoles adorally, there also exist biogeographic differences between the Zechstein basin and the tropical Tethyan carbonates of northern Italy (Mei & Henderson, 2001).

Description. This taxon is represented only from disarticulated spines (Figs. 6H–6I). These spines are striated and peg-like (up to 0.7 mm wide) and taper distally. They lack any visible ornamentation such as spinules and are widest where the spine meets the base. The acetabulum is round and perforated and they lack a milled ring (Figs. 6H–6I).

Remarks. Disarticulated spines recovered from washed residues display a morphology that is inconsistent with identification as crown group echinoids in the Palaeozoic or Early Mesozoic. These spines lack a milled ring (Figs. 6H–6I), a character which is found in many regular crown group echinoids, and appears to have first evolved in archaeocidarid echinoids during the Late Palaeozoic (Thompson et al., in press). Permian taxa with primary spines lacking a milled ring include Pronechinus cretensis König, 1982, Pronechinus anatoliensis Kier, 1965 both members of the Proterocidaridae, and Meekechinus elegans Jackson, 1912 of the Lepidesthidae. The stem group echinoid Yunnanechinus loupingensis Thompson et al., 2018 from the Anisian of Yunnan province, China and Lazarechinus mirabeti Hagdorn, 2018 from the Anisian of the French Muschelkalk (Thuy, Hagdorn & Gale, 2017) also display spines lacking a milled ring. All Permian crown group echinoids have spines with a milled ring (Kier, 1965; Smith & Hollingworth, 1990; Thompson et al., 2015; Thompson, Petsios & Bottjer, 2017) and this feature is also found in Permian archaeocidarids (Kittl, 1904; Boos, 1929; Kier, 1958b; Luepke, 1976; Simpson, 1976; Thompson, Petsios & Bottjer, 2017) and Lower and Middle Triassic cidaroids (Kier, 1968b; Kier, 1977; Hagdorn, 1995).

Because other ossicles representing stem group echinoid coronal plates were not recovered from washed residues, however, we note that it is necessary to treat the stem group nature of this material with caution.

Material. Three disarticulated rotulae (MPL 8654-1, MPL 8657-8, and MPL 8658-1), and three disarticulated hemipyramids (MPL 8651-2, MPL 8658-8, and MPL 8657-1). Furthermore, we include numerous fragmentary spines (MPL 8657-2, MPL 8654-3, MPL 8658-4, MPL 8653-4, MPL 8658-2, MPL 8658-7) whose taxonomic affinities are indeterminate.

Description. This material consists of numerous disarticulated elements of the Aristotle’s lantern and fragmentary spines which are indeterminate. The lantern elements recovered include rotulae and hemipyramids (Figs. 4A, 6A–6G). Three rotulae were recovered, which are shown in Figs. 6A–6E. The rotulae are bilaterally symmetrical, dorso-ventrally flattened, roughly rectangular in shape, and ranging from about two to five mm in length. All rotulae collected are hinge type (Märkel, 1979). Two morphotypes are known, hereby referred to as morphotypes A (Figs. 6A–6C) and B (Figs. 6D–6E). In photographs, all rotulae are oriented with the adaxial end (the end closest to the center of the lantern) oriented towards the bottom of the page.

Mophotype A is about equally wide throughout its width, except for the abaxial end, which narrows slightly passing from the epicondyles to the condyles. The dorsal surface is smooth throughout most of its length, with the exception of the abaxial end which ends in two slight bulges, which are the condyles. At the adaxial end of the dorsal surface, there is a shallow depression (Figs. 6A, 6C) in which the adaxial end of the compass would have rested. On the ventral surface of the rotula, slight indentations representing fossae are present at the adaxial end. At the abaxial end, the narrowing the condyles passing adaxially is clear (Fig. 6B) The abraded nature of the material makes it difficult to determine the location of rotular muscle scars.

Morphotype B is much wider at its adaxial end than abaxial end (Figs. 6D–6E). The dorsal side is smooth, and the condyles not very distinct. At the adaxial end, there is a small indentation, where the compass would rest. Also at the adaxial end, there is a slight constriction. On the ventral side, the condyles and epicondyles are much clearer (Fig. 6E) and two lateral bulges toward the adaxial end are the middle and inner fossae. Between these fossae are small indentations, which probably represent the rotular muscle scars.

Hemipyramids are small, no more than four mm tall. The foramen magnum is shallow, no more than 1/5 the height of the pyramid (Figs. 6F, 6G). A thickened wing edge is present along the edge of the hemipyramids (Fig. 6F) though this is only obvious in specimen MPL 8658-8.

The spines included herein are finely striate or covered in small rough spinules (Figs. 5M–5R). They range in length from about two mm to longer than five mm in length. They are circular in cross section. The spines are straight or show slight curvature in their orientation and many appear as small cylinders (Figs. 5N, 5P, 5R).

Remarks. The two rotulae morphotypes described herein; morphotype A and morphotype B, may represent variation within a single species, or, more likely, may represent rotulae from two different taxa. The two morphotypes differ most obviously in their general shape, with the adaxial end of morphotype B being much larger than that of A (compare Figs. 6A–6C to Figs. 6D–6E). Additionally in morphotype A, it is difficult to distinguish the precise location and number of fossae, while in morphotype B, the two symmetrical bulges located towards the adaxial end of the ventral side likely represent the inner and middle fossae. Insomuch as both of these morphotypes are hinge-type, they are similar to other rotulae reported from the Palaeozoic and Triassic. The rotula figured by Lewis & Ensom (1982) from the Mississippian Archaeocidaris whatleyensis Lewis & Ensom, 1982 is similar in its general rectangular shape, and in having a hinge joint, to the rotulae described herein. Furthermore, the rotula of the Anisian–Ladinian Triadotiaris grandaeva (v. Alberti, 1834), is similar to our morphotype 2 in widening adaxially (Hagdorn, 1995, fig. 5).

Hoare & Sturgeon (1976) figured three different rotulae in their study of echinoid ossicles from the Pennsylvanian Vanport Limestone of Ohio. Our morphotype B is similar to their figures 1 and 2 from their plate 2, while our morphotype A is similar to their plate 2, figure 3, albeit less narrow. In their study they attributed their disarticulated ossicles to two taxa, an archaeocidarid and a lepidesthid. Though the family level differences in rotula morphology are not well-known in stem group echinoids, it is possible that our two morphotypes represent a stem group and early crown group echinoid. Determining if this is the case is further complicated, however, in that all stem group echinoids where the rotulae are well-known, including Archaeocidaris M’Coy, 1844, Nortonechinus Thomas, 1920, and Polytaxicidaris Kier, 1958a and early crown group echinoids, such as T. grandaeva have hinge-type lanterns (Thomas, 1921; Kier, 1965; Märkel, 1979; Lewis & Ensom, 1982; Hagdorn, 1995). Further sampling of specimens where detailed morphology of rotulae, and of plates that are taxonomically informative at the family level (Thompson & Denayer, 2017), will be necessary to tie the details of differences in rotulae morphology to higher-taxonomic level classification in stem group echinoids.

The hemipyramids described herein are similar to those drawn by Smith & Hollingworth (1990) for E. keyserlingi. Both lanterns have shallow foramen magna, as is seen in crown group cidaroids. They are also much smaller than other hemipyramids potentially belonging to cidaroids described from the Capitanian of Texas (Thompson, Petsios & Bottjer, 2017). While the hemipyramids described by Thompson, Petsios & Bottjer (2017) were much too large to be attributed to the co-occurring cidaroid E. guadalupensis, the hemipyramids described herein are not much taller than a few millimeters, and could well belong to E. teseroensis. Though this may be the case, the incomplete nature of these hemipyramids and the fact that stem group echinoids likely also occur from the Tesero member precludes definitive attribution of these hemipyramids to a particular taxon.

The spines included herein are indeterminate given the state of their preservation, and could belong to any of the discussed taxa herein. Given the length and cylindrical nature of these spines, they may belong to Eotiaris teseroensis. Given their incomplete nature, however, we refrain from assigning them formally to that taxon.

Palaeoecological setting of the Tesero echinoids

The echinoid fauna reported herein is remarkable in that it occurs within the end-Permian extinction interval, about 2 m above the horizon representing the onset and peak of the EPME and about 0.4 m below the disappearance of the last stenotopic marine taxa (e.g., calcareous algae; e.g., Groves et al., 2007; Posenato, 2010; Fig. 2). It thus provides a glimpse into echinoid faunas and their palaeoecological setting in the immediate aftermath of the onset of extinction. Though this echinoid fauna occurs stratigraphically above the main extinction horizon, the occurrence of echinoids likely straddles the PTB, occurring in both the latest Permian and earliest Triassic (Fig. 2). The Tesero Member at the Tesero Section is characterized by basal oolitic beds, followed by an interval of oolitic and bioclastic grainstones and packstones interbedded with marly lenses, mudstones and microbialites (Posenato, 2009). The stratigraphic placement of the echinoid bearing samples is shown in Fig. 2. Two of the samples are from beds interpreted as microbial mats and represent shallow-water settings (CNT10 and CNT11; Noé, 1987; Posenato, 2009). CNT10 contains a fauna dominated by the brachiopod Orbicoelia (=?Crurithyris of Broglio Loriga et al., 1988), while CNT11 is dominated by Teserina nerii (Posenato, 2011). These shallow water brachiopod dominated assemblages in the Tesero section, despite occurring after the onset of the extinction event, yield faunal components characteristic of the Permian. Coeval benthic assemblages from a deeper-water setting (e.g., the Bulla section) are dominated by the bivalve Eumorphotis, and show a taxonomic composition similar to those occurring in the Early Triassic, when only a few eurytopic taxa were able to survive in heavily stressed marine conditions (e.g., Posenato, Holmer & Prinoth, 2014; Foster et al., 2017). The assemblages of the echinoid-bearing “mixed fauna” beds at the Tesero section thus record an ephemeral shallow water refuge for marine stenotopic taxa (Broglio Loriga, Neri & Posenato, 1986; Posenato, 2009) (Fig. 2). Echinoid remains are also known from the underlying Changhsingian Bulla Member of the Bellerophon Formation (Broglio Loriga et al., 1988; Posenato, 2009), though their taxonomic affinities are not well characterized.

In the western Dolomites (including the Tesero section), the extinction of stenotopic marine organisms occurred within a stratigraphic interval in the Tesero Member ranging from the contact with the underlying Bulla Member to the lower H. parvus Zone. The lower limit corresponds to the basal centimeters of the Tesero Member, which records the onset and peak of the extinction event, represented by a dramatic drop in fossil abundance and the loss of most foraminifers and large brachiopods (Posenato, 2010). The upper limit of the extinction interval is located within the lower H. parvus Zone, where the last stenotopic taxa occur rarely (Farabegoli, Perri & Posenato, 2007; Posenato, 2009; Posenato, 2010). The echinoid bearing beds (CNT6-CNT11A) occur stratigraphically above the onset of extinction, indicating that, at least in the Tesero section, echinoids survived the onset of extinction and associated environmental change. The microbialite-rich interval from which echinoids were collected at the Tesero section has been correlated with Beds 27a-b of the Meishan Section in South China (Farabegoli, Perri & Posenato, 2007). The stratigraphically lower echinoid bearing beds at Tesero (CNT6, CNT8, and CNT10) and beds 27a-b at Meishan are within the H. changxingensis Zone (Jiang et al., 2007; Lai, Jiang & Wignall, 2018) which is located within the “Survival Interval” at Meishan, where relict Permian taxa survived for ∼60,000 years up to the I. staeschei Zone (Shen et al., 2018). This correlation, together with the co-occurrence of echinoids and the “mixed” faunal assemblages at Tesero (Posenato, 2009), seems to indicate that the Tesero echinoids represent holdover taxa from Permian pre-extinction faunas. The delicate nature of the echinoid test recovered (Fig. 3), the occurrence of these echinoids in a microbialite, and the occurrence of articulated brachiopods (Broglio Loriga, Neri & Posenato, 1986), indicates that these taxa were not reworked from older strata and represent an autochthonous assemblage.

Echinoids and the end-Permian mass extinction